Respiratory syncytial virus (RSV) is an enveloped RNA virus that is the most important viral agent of pediatric respiratory tract disease worldwide. It currently lacks a vaccine or effective antiviral therapy. RSV is in Family Paramyxoviridae of the nonsegmented negative strand RNA viruses (mononegaviruses), and is the most complex member of the group. We previously determined the complete sequence of RSV genomic RNA and mapped its genes. Previous work included identification of transcription signals that begin and end each gene, the gene-start and gene-end signals, respectively, as well as the variable intergenic regions that lie between the genes. More recently, we have been examining additional features of the RSV genome to investigate possible roles or significance in RSV replication. For example, mutational analysis was performed on the gene-end signal of the NS1 or NS2 genes, or both, to investigate the importance of natural variation in this semi-conserved signal, and also to investigate the effect of individual signals on sequential transcription. This showed that the sequence differences that occur in nature indeed affect the efficiency of termination and polyadenylation. However, when these changes were introduced into infectious recombinant virus, they had little or no effect on sequential transcription or on virus replication in vitro or in the respiratory tract of mice, indicating that these variable nucleotides likely do not represent important inter-strain differences. We also uncoupled the gene overlap between the M2 and L gene, which had little effect on virus growth. Similarly, we uncoupled the M2-1 and M2-2 open reading frames so that they would be expressed by individual mRNAs rather than the natural mode of stop-restart translation from a single mRNA. This also had little effect on virus growth. Thus, these features of RSV genome organization are not critical to efficient replication and gene expression. We also mapped the major viral promoter using a cDNA-encoded minireplicon that is expressed intracellularly from transfected plasmids and complemented by plasmid-expressed viral proteins. This represents a system that is more amenable to large scale mutagenesis than is complete recombinant virus. Saturation mutagenesis of the first 26 nucleotides at the 3' end showed that the core polymerase for transcription and RNA replication involves the first 11 nucleotides of the genome, although additional downstream positions also have an effect. Interestingly, the positions involved in transcription versus RNA replication overlap but are not completely identical. In addition, transcription, but not replication, requires that the first gene-start signal be spaced correctly relative to the 3' end, implying that this might serve as part of the transcription promoter. These observations suggested that transcription, like RNA replication, involves polymerase contact at the very 3' end of the genome, but the differences in the promoters suggest that the two processes otherwise are distinct rather than interconvertible as had previously been thought. We also observed evidence of RNA recombination occurring in cell culture between two variants of RSV. Each of these input variants was a cDNA-derived virus that had been engineered to lack one or more viral genes and to contain several point mutations, thus providing markers for recombination. In this setting, a virus that undergoes recombination to acquire the full complement of viral genes would have a growth advantage. One such recombinant indeed was detected from the progeny of a coinfection, indicating that recombination can occur, likely by polymerase jumping, but that its frequency is low.